Stereoscopic warehouse multi-machine cooperative scheduling control method based on digital twinning

Abstract

The invention discloses a stereoscopic warehouse multi-machine cooperative scheduling control method based on digital twinning, and relates to the technical field of intelligent warehousing, and the method comprises the steps: collecting stereoscopic warehouse three-dimensional layout and multi-machine sensing data, and obtaining a standardized data stream through consistency verification; binding an instruction of the compliance control instruction set with target equipment by adopting a time sequence coding method, and issuing the instruction to the target equipment to obtain an execution feedback and backspacing mark; and according to the execution feedback and the rollback mark, performing fusion and noise reduction processing to form a continuously optimized scheduling cycle. According to the method, the standardized data flow is processed by adopting a multi-source feature fusion method, vibration, temperature and current features are integrated, and a healthy feasible region is generated, so that comprehensive evaluation and monitoring of the state of the warehouse equipment are realized, a scheduling framework for dynamic adjustment is provided, the resource utilization rate is optimized, and the scheduling accuracy and adaptability are improved.

Description

A Multi-Machine Collaborative Scheduling and Control Method for Automated Warehouses Based on Digital Twins Technical Field

[0001] This invention relates to the field of intelligent warehousing technology, and in particular to a multi-machine collaborative scheduling and control method for automated warehouses based on digital twins. Background Technology

[0002] In modern logistics and intelligent manufacturing systems, automated warehouses, as the core unit for high-density storage and efficient material flow, increasingly rely on information and automation technologies for multi-machine collaborative scheduling and control. Existing scheduling methods are usually based on static task planning and preset path rules, combined with local sensor feedback for equipment control, in order to achieve basic operation coordination and collision avoidance functions and support the stable operation of automated warehouses.

[0003] However, conventional methods have limitations in the integration of dynamic environmental adaptability and equipment health status scheduling: on the one hand, scheduling strategies often fail to fully integrate the real-time operating status of equipment, making it difficult to construct health constraint boundaries that evolve synchronously with physical entities; on the other hand, the lack of a continuous optimization mechanism based on closed-loop feedback leads to slot conflicts or resource contention in high-concurrency, high-density operation scenarios, affecting overall scheduling coordination. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a multi-machine collaborative scheduling and control method for automated warehouses based on digital twins to solve the problem of resource scheduling and collaborative optimization.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides a multi-machine collaborative scheduling and control method for automated warehouses based on digital twins. The method includes: collecting 3D layout data and multi-machine sensor data of the automated warehouse; obtaining a standardized data stream through consistency verification; processing the standardized data stream using a multi-source feature fusion method and integrating its features to generate a healthy feasible domain; generating equipment token reservations based on the healthy feasible domain by setting equipment time for intersections, elevator shafts, and grid fabric; transforming the equipment token reservations using a multi-objective method to obtain candidate control instructions, which are then verified for consistency to obtain a compliant control instruction set; binding the instructions of the compliant control instruction set to target equipment using a time-series encoding method and sending them to the target equipment to obtain execution feedback and rollback flags; and performing fusion and noise reduction processing based on the execution feedback and rollback flags to form a continuously optimized scheduling loop.

[0008] As a preferred embodiment of the digital twin-based multi-machine collaborative scheduling and control method for automated warehouses described in this invention, the specific steps for acquiring the three-dimensional layout of the automated warehouse and multi-machine sensor data, and obtaining a standardized data stream through consistency verification, are as follows:

[0009] Based on the three-dimensional layout of the automated warehouse and multi-machine sensor data, synchronous data acquisition is performed to obtain raw multi-source data packets;

[0010] A time alignment and coordinate unification method is used to perform unified timestamp and coordinate transformation on the original multi-source data packets to obtain a preliminary aligned dataset.

[0011] Perform consistency verification on the initially aligned dataset to obtain a standardized data stream.

[0012] As a preferred embodiment of the multi-machine collaborative scheduling and control method for automated warehouses based on digital twins described in this invention, the multi-source feature fusion method refers to aligning heterogeneous features from different sensors and data sources in time and space, and then processing them through denoising, weighting, and unified scaling to form a comprehensive feature representation.

[0013] As a preferred embodiment of the multi-machine collaborative scheduling and control method for automated warehouses based on digital twins described in this invention, the step of processing standardized data streams using a multi-source feature fusion method and integrating the features of the processed standardized data streams to generate a healthy feasible domain includes the following specific steps.

[0014] Based on standardized data streams, a multi-source feature fusion method is used to align vibration, temperature and current sequences and screen for segment stability to obtain a clean dataset.

[0015] Based on the cleaning dataset, a set of capability features is obtained by statistically integrating and correlating the vibration peak value, temperature change rate and current amplitude-frequency characteristics.

[0016] By integrating the set of capability features, a healthy feasible domain for scheduling constraints is generated.

[0017] As a preferred embodiment of the digital twin-based multi-machine collaborative scheduling and control method for automated warehouses described in this invention, the step of generating equipment token reservations by setting equipment time for intersections, elevator shafts, and grid fabric based on the health feasible domain is as follows:

[0018] Based on the healthy feasible region, a comparative analysis method is used to scan the occupancy range of intersections, risers and risers and key grids to obtain a candidate set of passage time.

[0019] Based on the candidate set of passage times, the entry, occupation and exit times of equipment are sorted in order to obtain the equipment passage time schedule;

[0020] By using the device access schedule and consistency verification, the time slots occupied by each device at the corresponding location are marked, and device token reservations are generated.

[0021] As a preferred embodiment of the multi-machine collaborative scheduling and control method for automated warehouses based on digital twins described in this invention, the candidate control instructions refer to a set of instructions to be checked and fine-tuned, generated based on a unified feasible domain, task priority, and token reservation.

[0022] As a preferred embodiment of the digital twin-based multi-machine collaborative scheduling and control method for automated warehouses described in this invention, the step of using a multi-objective method to convert equipment token reservations into candidate control instructions and passing consistency verification to obtain a compliant control instruction set is as follows:

[0023] Based on device token reservation, a multi-objective method is used to jointly solve the problem and obtain an initial set of candidate control commands.

[0024] Based on the initial set of candidate control instructions, the consistency between passage occupancy and token timing is verified to obtain the verified set of candidate control instructions.

[0025] By refining the verified candidate control instruction set, a compliance control instruction set is obtained.

[0026] As a preferred embodiment of the multi-machine collaborative scheduling and control method for automated warehouses based on digital twins described in this invention, the target device refers to the specific execution object that is explicitly assigned a path, time slot, station, and operation target by the compliant control instruction set during the automated warehouse scheduling process, and is subject to unified feasible domain and token reservation constraints.

[0027] As a preferred embodiment of the multi-machine collaborative scheduling and control method for automated warehouses based on digital twins described in this invention, the step of binding the instructions of the compliance control instruction set with the target device using a time-series encoding method and sending them to the target device to obtain execution feedback and rollback flags is as follows:

[0028] Based on the compliance control instruction set, a time-series encoding method is used for serialization and expiration marking to obtain an encoded instruction stream with timestamps and verification identifiers;

[0029] Based on the encoded instruction stream with timestamps and verification identifiers, the device identification information is associated with the channel occupancy status to generate the instruction packet for the target device and the access window;

[0030] Based on the instruction packets issued by the target device and the access window, execution feedback and rollback markers are obtained through rapid verification at the edge control terminal and on-site execution monitoring.

[0031] As a preferred embodiment of the multi-machine collaborative scheduling and control method for automated warehouses based on digital twins described in this invention, the step of performing fusion and noise reduction processing based on execution feedback and rollback flags to form a continuously optimized scheduling loop includes the following specific steps.

[0032] Based on execution feedback and rollback markers, the 3D layout of the automated warehouse and multi-machine sensor data are fused and denoised to obtain a cleaned execution dataset.

[0033] Based on the cleaned execution dataset, comparative analysis is used to assess and locate health boundary offset, token default distribution and slot congestion hotspots, and to generate adjustment suggestions for the healthy feasible domain.

[0034] Based on the adjustment recommendations, online updates are performed and the unified healthy feasible domain scheduling weights are optimized synchronously to form a continuously optimized scheduling loop.

[0035] The beneficial effects of this invention are as follows: By processing standardized data streams using a multi-source feature fusion method and integrating vibration, temperature, and current characteristics to generate a healthy feasible domain, a comprehensive assessment and monitoring of warehouse equipment status is achieved, providing a dynamically adjustable scheduling framework, optimizing resource utilization, and improving the accuracy and adaptability of scheduling; by using a time-series coding method to encode and timestamp compliance control instructions, binding instructions to target equipment and executing them, the timeliness and accuracy of instructions are effectively enhanced, achieving efficient collaboration and precise scheduling of the warehouse, and improving the level of intelligence in warehouse management and resource utilization efficiency. Attached Figure Description

[0036] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 is a flowchart of a multi-machine collaborative scheduling and control method for automated warehouses based on digital twins.

[0038] Figure 2 is a flowchart of data processing and health feasible domain generation.

[0039] Figure 3 is a flowchart of device token reservation generation.

[0040] Figure 4 is a flowchart of the generation and execution of compliance control instruction sets. Detailed Implementation

[0041] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0042] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0043] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0044] Referring to Figures 1-4, an embodiment of the present invention is provided, which offers a multi-machine collaborative scheduling and control method for an automated warehouse based on digital twins, comprising the following steps:

[0045] S1: Collect 3D layout data and multi-machine sensor data of the automated warehouse, and obtain a standardized data stream through consistency verification.

[0046] S1.1: Based on the three-dimensional layout of the automated warehouse and multi-machine sensor data, synchronous data acquisition is performed to obtain raw multi-source data packets.

[0047] Furthermore, based on the three-dimensional layout of the automated warehouse, the acquisition points, equipment identification, and spatial reference information of vibration sensors, temperature sensors, and current sensors are determined, and multi-machine sensor data acquisition terminals are synchronously acquired. During the synchronous acquisition process, each record is supplemented with equipment identification, acquisition time stamp, and coordinate reference description from the three-dimensional layout of the automated warehouse (which means uniformly converting the three-dimensional layout of the automated warehouse and multi-machine sensor data to the same spatial coordinate system to achieve consistent expression of spatial location), and is continuously encapsulated and sequentially numbered according to equipment identification and channel division to obtain the original multi-source data packet.

[0048] Specifically, the three-dimensional layout of an automated warehouse refers to the overall structural arrangement that uses a unified three-dimensional layout coordinate system to accurately express the position and height relationship of shelves, aisles, elevators, conveyor lines and operating equipment in space.

[0049] S1.2: Using a time alignment and coordinate unification method, the original multi-source data packets are transformed with unified timestamps and coordinates to obtain a preliminary aligned dataset.

[0050] Furthermore, a time alignment and coordinate unification method is adopted to perform unified timestamp processing on the original multi-source data packets. The time alignment strategy is used to synchronize the vibration sequence, temperature sequence, and current sequence on a unified time axis. Based on the coordinate reference provided by the three-dimensional layout of the automated warehouse, coordinate transformation is completed to express the equipment acquisition pose and measurement point position in a unified coordinate manner. After unified timestamp and coordinate transformation, the data records are organized and consistency verified according to equipment identification, sampling order, and spatial index to eliminate time sequence conflicts and spatial duplication, and obtain a preliminary aligned dataset.

[0051] It should be noted that the unified timestamp processing is based on the time standardization and alignment of multi-machine sensor records using a reference time source, mapping vibration sequences, temperature sequences and current sequences to the same time axis for synchronous analysis and consistency verification.

[0052] Consistency verification refers to comparing and verifying the timing, coordinates, access occupancy, and token timing of the preliminary aligned dataset and the candidate control instruction set under a unified time base and spatial reference, to obtain a standardized data stream and a verified candidate control instruction set and compliant control instruction set.

[0053] The timing alignment strategy refers to resampling and interpolation registration of vibration sequences, temperature sequences and current sequences based on unified timestamps and clock synchronization, so that multi-source data packets are mapped to the same time axis for synchronous processing.

[0054] Specifically, the time alignment and coordinate unification method refers to performing unified timestamps and coordinate transformations on the original multi-source data packets, so that vibration sequences, temperature sequences and current sequences are aligned under the same time base and spatial reference, and then statistically integrated and verified for consistency.

[0055] It should be noted that the vibration sequence, temperature sequence, and current sequence are obtained from the synchronously acquired multi-machine sensor data through time alignment and coordinate unification methods to achieve unified timestamps and coordinate transformation. After anomalies are eliminated through consistency verification, they are obtained from the standardized data stream by sensor channel deconstruction and time window slicing.

[0056] S1.3: Perform consistency verification on the initially aligned dataset to obtain a standardized data stream.

[0057] Furthermore, based on the 3D layout of the automated warehouse, the location markers and coordinate directions in the initially aligned dataset are compared to verify the continuity and uniqueness of the timestamps after time alignment. Statistical methods are used to verify the uniformity of vibration, temperature, and current in the initially aligned dataset, merging duplicate records and correcting outliers to improve record integrity. Simultaneous location and event comparisons between multi-machine sensor data confirm cross-source consistency and temporal consistency. Based on the consistency verification results, field standardization, coordinate unification, and timestamp unification are completed to obtain a standardized data stream.

[0058] S2: A multi-source feature fusion method is used to process the standardized data stream, and the processed standardized data stream is then integrated to generate a healthy feasible domain.

[0059] S2.1: Multi-source feature fusion method refers to the method of aligning heterogeneous features from different sensors and data sources in time and space, and forming a comprehensive feature representation after denoising, weighting and unifying scale processing.

[0060] A superior multi-source feature fusion method, under a unified time axis and a unified coordinate reference, performs resampling and interpolation registration, segment stability screening and consistency verification on vibration sequences, temperature sequences and current sequences, ensuring the temporal consistency and spatial correspondence of the capability feature set, and providing a synchronous constraint basis for health feasible domain and equipment token reservation.

[0061] S2.2: Based on standardized data streams, a multi-source feature fusion method is used to align vibration, temperature and current sequences and screen for segment stability to obtain a clean dataset.

[0062] Furthermore, based on standardized data streams, time alignment and coordinate unification are performed on vibration, temperature, and current sequences, and comparative analysis is conducted under a unified time axis and coordinate reference. Multi-source feature fusion methods are used for feature denoising and unified healthy feasible domain scheduling weight integration, enabling vibration, temperature, and current sequences to be represented on a unified scale. Consistency checks are performed to verify continuity, volatility, and amplitude-frequency correlation, marking stable segments on a unified time axis. Usable segments are formed by the intersection of these stable segments. Based on the distribution of vibration, temperature, and current characteristics within a given time window, segments that do not meet stability conditions are filtered out, retaining only valid data. After completing the stable segment filtering, a clean dataset is obtained.

[0063] It should be noted that comparative analysis refers to comparing the peak co-occurrence, checking the rate of change in the same direction, and comparing the amplitude and frequency cross-sections of vibration, temperature, and current sequences to confirm stable segments that are mutually corroborated in time and have consistent coordinates.

[0064] The unified healthy feasible domain scheduling weight integration is based on time alignment, coordinate unification and segment stability screening. It performs noise suppression on vibration sequence, temperature sequence and current sequence and performs statistical integration and correlation correction on a unified scale.

[0065] Segment stability screening refers to performing peak co-occurrence comparison, rate of change unidirectional test, and amplitude-frequency comparison on vibration, temperature, and current sequences after time alignment and coordinate unification. Combined with consistency verification, stable segments are marked on a unified time axis, and usable segments are formed by the intersection of stable segments.

[0066] S2.3: Based on the clean dataset, a set of capability features is obtained by statistically integrating and correlating the vibration peak value, temperature change rate, and current amplitude-frequency characteristics.

[0067] Furthermore, based on the clean dataset, peak detection and noise suppression are performed on the vibration sequence to extract the vibration peak value, and frequency domain analysis is performed on the temperature and current sequences to form the current amplitude-frequency characteristics. Scale unification and statistical extraction are carried out among the vibration peak value, temperature change rate and current amplitude-frequency characteristics, and statistical integration and correlation correction are performed to eliminate the offset caused by alignment residuals. The capability feature set is obtained by integrating the unified health and feasible domain scheduling weights (the basis for statistical integration and correlation correction comes from the joint statistics and consistency verification output of the clean dataset).

[0068] It should be noted that noise suppression refers to the process of locating extreme values ​​of vibration sequences under a unified time axis and coordinate reference, based on multi-source feature fusion methods and segment stability screening, and eliminating discontinuous fluctuations and interferences through consistency verification, in order to obtain a stable vibration peak characterization.

[0069] The statistical extraction process involves mapping the values ​​of vibration peak value, temperature change rate and current amplitude-frequency characteristics to a unified scale by performing unit unification and standardization transformation, and then calculating the statistical measures of mean, variance, skewness, kurtosis and quantile on the unified scale to complete the statistical extraction.

[0070] Specifically, the vibration peak value, temperature change rate, and current amplitude-frequency characteristics are based on a clean dataset. A multi-source feature fusion method is used to perform statistical integration and correlation correction on a unified scale. The vibration peak value is extracted from the vibration sequence, the temperature change rate is calculated from the temperature sequence, and the current amplitude-frequency characteristics are characterized from the current sequence.

[0071] Correlation correction refers to extracting vibration peaks from vibration sequences through peak detection, obtaining temperature change rates from temperature sequences, and characterizing current amplitude-frequency characteristics from current sequences through spectral analysis under a unified time axis and standardized scale, and using the statistics between them to correct outliers, phase shifts, and scale deviations, forming a consistent and comparable set of capability characteristics.

[0072] Frequency domain analysis refers to the process in multi-source feature fusion methods where vibration sequences selected based on time alignment and segment stability are combined with consistency verification, statistical integration, and correlation correction to differentiate adjacent time points and suppress abnormal fluctuations in order to form a stable temperature change rate characterization.

[0073] S2.4: By integrating the capability feature set, a healthy feasible domain for scheduling constraints is generated.

[0074] Furthermore, based on the capability feature set, the vibration peak value, temperature change rate and current amplitude-frequency characteristics are standardized, and a health metric field is formed by integrating the unified health feasible domain scheduling weights. The current amplitude-frequency characteristics of abnormal fluctuation sections are integrated by consistency verification, and continuous usable intervals are obtained by section stability screening. On a unified coordinate and time axis, the capability feature set is processed by correlation correction to generate a health feasible domain with scheduling constraints.

[0075] It should be noted that the abnormal fluctuation segment refers to a continuous time segment in which the vibration peak value, temperature change rate, and current amplitude-frequency characteristics show abrupt changes, jumps, and non-periodic shifts relative to their respective stable baselines under a unified time axis and coordinate system.

[0076] Current amplitude-frequency characteristic integration refers to the statistical integration and correlation correction of the capability feature set under a unified scale in the multi-source feature fusion method, and the generation of a healthy feasible domain for scheduling constraints by combining unified healthy feasible domain scheduling weight integration and consistency verification.

[0077] Specifically, scheduling constraints are used to limit the occupancy conditions and priority relationships of device paths, time slots, positions and speeds on a unified coordinate and time axis, ensuring the consistency and executability of device token reservation and compliance control instruction set generation.

[0078] S3: Based on the health and feasibility domain, generate equipment token reservations by setting equipment time for intersections, lifts, and grid fabrics.

[0079] S3.1: Based on the healthy feasible domain, a comparative analysis method is used to scan the occupancy intervals of intersections, risers and risers and key grids to obtain a candidate set of passage times.

[0080] Furthermore, based on the healthy feasible domain, available time slices and space occupancy constraints are extracted from intersections, elevator shafts, and key grids. A comparative analysis method is used to compare available time slices at different locations one by one to identify time segments that simultaneously meet the requirements of non-conflict and continuity. The time slices are then merged, deduplicated, and ordered. After cross-location alignment is completed, consistency verification is performed based on the healthy feasible domain to eliminate time slices that do not match the space occupancy constraints, thus forming a candidate set of passage times.

[0081] It should be noted that by performing time alignment and coordinate unification on the position status sequences of intersections, elevator shafts and key grids within the healthy feasible domain, and by combining the consistency verification of the 3D layout of the automated warehouse, equipment identification information and channel occupancy status, non-occupied continuous sections are located as occupancy time slices, and spatial occupancy constraints are generated based on the geometric boundaries of the 3D layout of the automated warehouse and the channel occupancy status.

[0082] The comparative analysis method refers to aligning the available time slices of intersections, risers and falls, and key grids with time and position under the constraints of the healthy feasible domain. By comparing the overlapping relationships of interval intersections and mergers, it identifies non-conflicting and continuous passage sections, and performs merging, deduplication, and ordering of the time slices.

[0083] Specifically, the healthy feasible domain constraint refers to determining the safe upper and lower limits and combined allowable ranges of each characteristic component within a unified characteristic space composed of vibration peak value, temperature change rate and current amplitude-frequency characteristics, based on the equipment rated parameters, historical normal operating condition statistical intervals and online feedback evaluation results, and using the state falling within the interval as the healthy feasible scheduling constraint condition.

[0084] A better approach is to combine consistency verification to eliminate time slices that do not match the occupancy constraints, thus generating a comparative result that can be used to generate a candidate set of passage times.

[0085] S3.2: Based on the candidate set of passage times, sort the entry, occupation and exit times of the equipment in sequence to obtain the equipment passage time schedule.

[0086] Furthermore, based on the candidate set of passage times, and according to the path order of the equipment at the intersection, the elevator shaft, and the grid cloth, continuous time windows of entry time, occupation time, and exit time are matched for each equipment. By identifying the overlap of the occupancy intervals of shared locations and performing sequential sorting, boundary alignment and position correspondence are performed on adjacent time windows of the same equipment to ensure the continuity of occupation. At the same time, time connection constraints are applied between different locations to obtain the equipment passage time schedule.

[0087] Specifically, time continuity constraints refer to limiting the sequence and continuity of start and end times and occupied time slots of the same path and related operations during the scheduling of multiple devices and multiple processes, so that the preceding and following processes do not overlap or conflict on a unified time axis, nor do they create time gaps that violate process safety requirements.

[0088] It should be noted that the overlap determination is made by comparing the start and end times of any two time windows at the shared location. If the maximum start time is earlier than the minimum end time, it is determined to be an overlap.

[0089] Specifically, shared locations are determined by analyzing the intersections, elevator shafts, and key grids of the equipment paths after unifying the 3D layout and coordinates of the automated warehouse, identifying the spaces that can be occupied by multiple devices simultaneously.

[0090] The unified coordinate device path refers to the path sequence on the grid cloth that transforms the path coordinates of the target device to the unified coordinate system of the three-dimensional layout of the automated warehouse through time alignment and coordinate unification methods, and represents it with a consistent index sequence.

[0091] S3.3: By using the device access schedule and consistency verification, mark the time slot occupied by each device at the corresponding location and generate device token reservation.

[0092] Furthermore, based on the equipment access schedule, the entry, occupancy, and exit time periods of each piece of equipment are retrieved one by one according to intersections, elevator shafts, and key grids. Consistency verification is used to check the continuity of time, mutual exclusion of resources, and consistency with the healthy feasible domain constraints to determine the order of priority and occupancy limits, so that the equipment access schedule is consistent with the healthy feasible domain. After completing the occupancy time slot verification, the occupancy time slot of each piece of equipment at the corresponding location is marked in the form of tokens, and the equipment token reservation is generated.

[0093] It should be noted that time slot verification involves verifying the consistency between the equipment access schedule and the health and feasibility domain, comparing the time continuity, resource exclusivity, and token timing of the entry, occupation, and exit periods of intersections, elevator shafts, and key grids, marking conflicts, and confirming available occupation time slots.

[0094] It should be noted that item-by-item retrieval refers to sequentially reading and comparing the relevant constraints for each device during its entry, occupancy, and exit times at intersections, elevator shafts, and key grids, according to the recorded order of the equipment passage schedule, to complete the verification and annotation of a single record.

[0095] S4: A multi-objective approach is used to transform device token reservations to obtain candidate control instructions, which are then verified for consistency to obtain a set of compliant control instructions. Candidate control instructions refer to the set of instructions to be verified and fine-tuned based on a unified feasible domain, task priority, and token reservation.

[0096] S4.1: Based on device token reservation, a multi-objective method is used to solve the problem jointly to obtain an initial set of candidate control commands.

[0097] Furthermore, based on device token reservation, decision variables such as path, time slot, speed, and station location are constructed and combined with space occupancy constraints, consistency verification results, and health feasible domain to form a constraint set. Objective functions for passage time, conflict risk, path connectivity, and execution stability are established and solved jointly using a multi-objective method. Through candidate solution feasibility screening and consistency verification, a spatiotemporal instruction sequence and action parameters matching device token reservation are generated and aggregated into an initial candidate control instruction set.

[0098] Specifically, the spatiotemporal instruction sequence and action parameters refer to a set of control instructions arranged in chronological order and bound to spatial location, including action parameters such as path, time slot, speed, and station position, which are used to drive the target device to execute sequentially under the constraints of device token reservation and health feasible domain.

[0099] The multi-objective method refers to the optimization objective that needs to be balanced between travel time, conflict risk, path connectivity and execution stability. It uses a general strategy of weighted summation and hierarchical serialization non-dominated sorting to jointly solve for the solution set that satisfies the healthy feasible region constraint and generate candidate control instructions.

[0100] S4.2: Based on the initial set of candidate control instructions, verify the consistency between access occupancy and token timing to obtain the verified set of candidate control instructions.

[0101] Furthermore, based on the initial set of candidate control instructions, a passage occupancy time slot table and a token timing table are constructed. The position sequence and time sequence of each candidate control instruction are compared. A comparative analysis method is used to detect occupancy conflicts and window overlaps at intersections, elevator shafts, and grid structures. Path continuity, entry and exit sequence, and consistency of equipment passage time schedules are verified, and the source and stage of conflict are marked. For candidate control instructions with conflicts, time slot adjustments and sequence rearrangements are performed to maintain consistency in equipment token reservations and prevent disruption of space occupancy constraints. After completing the occupancy consistency verification and token timing verification, a verified set of candidate control instructions that satisfies the consistency of passage occupancy and token timing is obtained.

[0102] Specifically, the process of constructing the passage occupancy time slot table and the token time sequence table involves parsing the path, entry, occupancy and exit time periods and corresponding equipment token reservations for each candidate control command based on the initial candidate control command set. The data is aggregated by intersection, elevator entrance and grid cloth and sorted by time and divided into continuous intervals. The passage occupancy time slot table with coverage location and time period is obtained by summarizing the data. The token time sequence table is then generated by concatenating the tokens according to the order of token issuance and duration.

[0103] S4.3: By refining the verified candidate control instruction set, a compliance control instruction set is obtained.

[0104] Furthermore, based on the verified candidate control instruction set, the path, time slot, speed, and station position fields are sorted and aligned according to task priority and equipment token reservation. Time slot fine-tuning and occupancy sequence rearrangement are performed for passage occupancy conflicts. Duplicate and redundant instructions are merged and deduplicated. Smoothing and continuity correction are performed based on the healthy and feasible domain to ensure that the candidate control instruction set is consistent with the equipment passage schedule, token sequence, and channel occupancy status, thus forming a compliant control instruction set.

[0105] It should be noted that merging and deduplication refers to aggregating the candidate control command sets after verification that the target device, path, time slot, and station location fields are consistent and can be continuously spliced, and deleting duplicate records to ensure the uniqueness of the commands and the compactness of the execution sequence.

[0106] S5: Use time-series encoding to bind the instructions of the compliance control instruction set to the target device and send them to the target device to obtain execution feedback and rollback flags.

[0107] S5.1: Target equipment refers to the specific execution object that is explicitly assigned a path, time slot, station and operation target by the compliance control instruction set during the scheduling of automated warehouses, and is subject to the constraints of unified feasible domain and token reservation.

[0108] S5.2: Based on the compliance control instruction set, a time-series encoding method is used for serialization and validity period marking to obtain an encoded instruction stream with timestamps and verification identifiers.

[0109] Furthermore, based on the compliance control instruction set, a time-series encoding method is used to organize each compliance control instruction into an ordered record according to a fixed field order. First, the record is serialized and marked with a generation time timestamp and an expiration time mark (serialization field position order and verification identifier generation rules). The records are then concatenated into a continuous byte stream according to time sequence and an order association between instructions is established to form an encoded instruction stream with timestamps and verification identifiers.

[0110] Specifically, each compliance control instruction and tag refers to the generation time timestamp and expiration time tag attached to the compliance control instruction, which are used to clarify the effective start and end dates and execution window and support serialization management in the time series encoding method.

[0111] It should be noted that time series coding refers to a coding method that serializes control commands and status data in chronological order, adds timestamps, sequence numbers and validity period identifiers to each time segment, and realizes the orderly parsing, scheduling and tracing of commands on a unified time axis.

[0112] Ideally, compliance control instructions should be issued in an orderly manner according to time, with clear validity periods, verifiable integrity, and traceability, thereby improving the reliability of issuance and the consistency of execution.

[0113] S5.3: Based on the encoded instruction stream with timestamps and verification identifiers, associate device identification information with channel occupancy status to generate a command packet for the target device and the access window.

[0114] Furthermore, based on the encoded instruction stream with timestamps and verification identifiers, the target device identifier, path, time slot, and station position fields are parsed. Using the device identification information as the key, time alignment and position matching are performed with the occupied time slot records of the channel occupancy status to determine the passage window corresponding to the target device. The matching results are then subjected to consistency verification based on the verification identifier and sorted by timestamp. The path, time slot, station position, passage window, and verification identifier are encapsulated to generate the instruction packet for the target device and the passage window.

[0115] It should be noted that location matching involves synchronously comparing the instruction time slot with the occupied time slot of the channel occupancy status according to the timestamp of the time series encoding method, and performing spatial matching based on the unified coordinate path and the index position of the station on the grid cloth to complete time alignment and location matching.

[0116] The instruction packet refers to the set of instructions that are sent by parsing the target device identifier, path, time slot and station position fields based on the encoded instruction stream with timestamp and check mark, combining the channel occupancy status to complete time alignment and position matching, and encapsulating the path, time slot, station position, passage window and check mark in order of timestamp.

[0117] Specifically, parsing the target device identifier, path, time slot, and station location fields is based on the encoded instruction stream of timestamps and verification identifiers. Syntax parsing and type verification are performed according to the field position order and delimiters of the time sequence encoding method, and the target device identifier, path, time slot, and station location fields are directly extracted from the instruction text.

[0118] S5.4: Based on the instruction packets issued by the target device and the access window, execution feedback and rollback flags are obtained through rapid verification at the edge control terminal and on-site execution monitoring.

[0119] Furthermore, based on the instruction packets issued to the target device and the access window, the edge control terminal performs integrity checks on the timestamp and verification identifier, and verifies the consistency between the target device identification information and the channel occupancy status. Once the verification is successful, the instruction is executed. On-site execution monitoring continuously collects the status sequence of the target device in terms of path, time slot, position, and speed based on multi-machine sensor data. The monitoring sequence is time-aligned and field-compared with the compliance control instruction set. Access window occupancy deviation, token default, and work completion status are recorded, and a rollback flag is registered according to the deviation type. At the same time, the process status and result status are summarized to form execution feedback, and the output execution feedback and rollback flag are obtained.

[0120] It should be noted that the edge control terminal refers to the computing and control entity deployed near the site of the automated warehouse that directly interacts with the target equipment and multi-machine sensor data.

[0121] A superior system can complete the verification of timestamps and verification identifiers, the consistency verification of target device identification information and channel occupancy status, and the on-site execution monitoring and command issuance.

[0122] S6: Based on the execution feedback and rollback flags, perform fusion and noise reduction processing to form a continuously optimized scheduling loop.

[0123] S6.1: Based on execution feedback and rollback markers, the 3D layout of the automated warehouse and multi-machine sensor data are fused and denoised to obtain a cleaned execution dataset.

[0124] Furthermore, based on execution feedback and backoff markers, the 3D layout of the automated warehouse and multi-machine sensor data are time-aligned and coordinate-unified to obtain the correspondence between execution feedback and spatial location. Backoff markers are used to annotate potentially mismatched segments, and fusion and noise reduction processing are employed to suppress noise in vibration, temperature, and current sequences. Then, consistency verification and channel occupancy status cross-checking are used to remove outliers and repair continuity gaps. After completing the statistical update, multi-source fusion is implemented using a unified healthy feasible domain scheduling weight integration method, while maintaining a consistent mapping with the 3D layout topology of the automated warehouse, to obtain a cleaned execution dataset.

[0125] It should be noted that the correspondence between execution feedback and spatial location is obtained by performing unified timestamp alignment and coordinate transformation on the execution feedback, matching the equipment identification information with the grid, intersection and aisle occupancy status in the three-dimensional layout of the automated warehouse, and determining the correspondence between execution feedback and spatial location within a unified coordinate system.

[0126] Rollback marking refers to the information marked on time slices, path segments, and passage windows that were not completed in accordance with the compliant control instruction set, based on the results of on-site execution monitoring and rapid verification at the edge control terminal during the execution of the target device.

[0127] Specifically, fusion and noise reduction processing refers to the time alignment and coordinate unification of the 3D layout of the automated warehouse and multi-machine sensor data under the constraints of execution feedback and rollback marking. Multi-source fusion is completed through consistency verification and unified health and feasible domain scheduling weight integration to obtain the cleaned execution dataset.

[0128] S6.2: Based on the cleaned execution dataset, use comparative analysis to assess and locate health boundary offset, token default distribution and slot congestion hotspots, and generate adjustment suggestions for the healthy feasible domain.

[0129] Furthermore, based on the cleaned execution dataset, a comparative analysis method is used to establish the correspondence between occupancy time slot sequences and token default records according to the dimensions of intersections, elevator shafts, and grid layout. The occupancy persistence, occupancy overlap, and token default frequency are grouped and statistically analyzed for different time periods and locations to obtain the dense areas of time slot sequences with health boundary offset. The comparative analysis method is used to compare the dense areas of occupancy time slot sequences from both spatial and temporal perspectives to locate time slot congestion hotspots and associated equipment passage sequences. An evaluation is conducted among health boundary offset, token default distribution, and time slot congestion hotspots to clarify the suggestions for shrinking and widening the occupancy range of intersections, elevator shafts, and grids. A synchronous revision path for integrating the equipment passage schedule and the unified health feasible domain scheduling weight is proposed to generate adjustment suggestions for the health feasible domain.

[0130] It should be noted that the assessment between health boundary offset, token default distribution and slot congestion hotspots refers to comparing the execution data with the historical baseline health feasible domain, respectively statistically analyzing the characteristic out-of-bounds offset of each device location, the default frequency and degree of advance and delay of token reservation and actual occupied slots, as well as the occupancy rate and queue length of the time grid, and comprehensively quantifying and ranking the three under a unified indicator system.

[0131] Specifically, the direction of the healthy feasible domain offset indicates the direction of the healthy feasible domain boundary's inward or outward movement relative to the previous cycle on the spatial or temporal axis, while the degree of offset indicates the magnitude of the healthy feasible domain boundary displacement after unit unification.

[0132] Time slot congestion hotspots refer to time intervals and location combinations in the time slot occupancy sequence of intersections, riser gates, and key grids where the passage demand of multiple devices overlaps and remains concentrated.

[0133] S6.3: Based on the adjustment suggestions, perform online updates and synchronous optimization of the unified healthy feasible domain scheduling weights to form a continuously optimized scheduling loop.

[0134] Furthermore, the adjustment suggestions output by the comparative analysis method are used to update the health feasible domain online. Combined with the consistency verification results, the occupied time slot label and task priority in the device token reservation are synchronized and unified to optimize the health feasible domain scheduling weight. The capability feature set generated by the multi-source feature fusion method participates in the unified health feasible domain scheduling weight integration under a unified scale. This allows the multi-objective method to use the optimized unified health feasible domain scheduling weight coefficient in real time during the joint solution of path, time slot, speed and station. The compliance control instruction set is re-encoded and issued using the time series encoding method. The execution feedback and rollback mark re-enter the comparative analysis method and consistency verification to form a closed loop, forming a continuously optimized scheduling cycle.

[0135] It should be noted that the basis for synchronously optimizing the scheduling weights of the unified healthy feasible domain is the grouping statistical results and consistency verification results of the comparative analysis method.

[0136] In summary, this invention achieves comprehensive assessment and monitoring of warehouse equipment status by: processing standardized data streams using a multi-source feature fusion method and integrating vibration, temperature, and current characteristics to generate a healthy feasible domain; providing a dynamically adjustable scheduling framework; optimizing resource utilization; and improving the accuracy and adaptability of scheduling. Furthermore, it employs a time-series coding method to encode and timestamp compliance control instructions, binding and executing these instructions to target equipment, effectively enhancing the timeliness and accuracy of instructions. This results in efficient warehouse collaboration and precise scheduling, improving the intelligence level of warehouse management and resource utilization efficiency.

[0137] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A multi-machine collaborative scheduling and control method for an automated warehouse based on digital twins, characterized in that: This includes collecting 3D layout data and multi-machine sensor data of the automated warehouse, and obtaining a standardized data stream through consistency verification; A multi-source feature fusion method is used to process the standardized data stream, and the processed standardized data stream is then integrated to generate a healthy feasible domain. Based on the healthy feasible domain, equipment token reservations are generated by setting equipment time for intersections, elevators, and grids. A multi-objective method is used to transform the equipment token reservations to obtain candidate control instructions, which are then verified for consistency to obtain a compliant control instruction set. A time-series encoding method is used to bind the instructions of the compliant control instruction set to the target equipment and send them to the target equipment to obtain execution feedback and rollback flags. Based on the execution feedback and rollback flags, fusion and noise reduction processing are performed to form a continuously optimized scheduling loop.

2. The method for multi-machine collaborative scheduling and control of an automated warehouse based on digital twins as described in claim 1, characterized in that: The process of collecting 3D layout and multi-machine sensor data of the automated warehouse and obtaining a standardized data stream through consistency verification is as follows: Based on the 3D layout and multi-machine sensor data of the automated warehouse, synchronous collection is performed to obtain the original multi-source data packets; the original multi-source data packets are uniformly timestamped and coordinate transformed using time alignment and coordinate unification methods to obtain a preliminary aligned dataset. Perform consistency verification on the initially aligned dataset to obtain a standardized data stream.

3. The method for multi-machine collaborative scheduling and control of an automated warehouse based on digital twins as described in claim 2, characterized in that: The multi-source feature fusion method refers to aligning heterogeneous features from different sensors and data sources in time and space, and then processing them through denoising, weighting, and unified scaling to form a comprehensive feature representation.

4. The method for multi-machine collaborative scheduling and control of an automated warehouse based on digital twins as described in claim 3, characterized in that: The process involves using a multi-source feature fusion method to process the standardized data stream and then integrating the features of the processed standardized data stream to generate a healthy feasible domain. The specific steps are as follows: Based on the standardized data stream, the multi-source feature fusion method is used to align the vibration, temperature, and current sequences and screen for segment stability to obtain a clean dataset. Based on the cleaning dataset, a set of capability features is obtained by statistically integrating and correlating the vibration peak value, temperature change rate and current amplitude-frequency characteristics. By integrating the set of capability features, a healthy feasible domain for scheduling constraints is generated.

5. The method for multi-machine collaborative scheduling and control of an automated warehouse based on digital twins as described in claim 4, characterized in that: The process of generating equipment token reservations based on the health feasible domain involves setting equipment time for intersections, elevator shafts, and grid structures. The specific steps are as follows: Based on the health feasible domain, a comparative analysis method is used to scan the occupancy intervals of intersections, elevator shafts, and key grid structures to obtain a candidate set of passage times; based on the candidate set of passage times, the entry, occupancy, and exit times of equipment are sequentially arranged to obtain a equipment passage timetable; using the equipment passage timetable, consistency verification is used to mark the occupancy time slots of each device at the corresponding location, generating an equipment token reservation.

6. The method for multi-machine collaborative scheduling and control of an automated warehouse based on digital twins as described in claim 5, characterized in that: The candidate control instructions refer to the set of instructions to be checked and fine-tuned, generated based on a unified feasible domain, task priority, and token reservation.

7. The method for multi-machine collaborative scheduling and control of an automated warehouse based on digital twins as described in claim 6, characterized in that: The process of transforming device token reservations using a multi-objective method to obtain candidate control instructions and passing consistency verification to obtain a compliant control instruction set involves the following steps: Based on device token reservations, a multi-objective method is used for joint solution to obtain an initial candidate control instruction set; based on the initial candidate control instruction set, consistency verification is performed on access occupancy and token timing to obtain a verified candidate control instruction set; and the verified candidate control instruction set is refined to obtain a compliant control instruction set.

8. The method for multi-machine collaborative scheduling and control of an automated warehouse based on digital twins as described in claim 7, characterized in that: The target equipment refers to the specific execution object that is explicitly assigned a path, time slot, station and operation target by the compliance control instruction set during the scheduling of the automated warehouse, and is subject to the constraints of the unified feasible domain and token reservation.

9. The method for multi-machine collaborative scheduling and control of an automated warehouse based on digital twins as described in claim 8, characterized in that: The process involves using time-series encoding to bind the instructions of the compliance control instruction set to the target device and then sending them to the target device to obtain execution feedback and rollback flags. The specific steps are as follows: Based on the compliance control instruction set, a time-series encoding method is used for serialization and validity period marking to obtain an encoded instruction stream with timestamps and verification identifiers; based on the encoded instruction stream with timestamps and verification identifiers, device identification information and channel occupancy status are associated to generate a command packet for the target device and the access window; based on the command packet for the target device and the access window, execution feedback and rollback flags are obtained through rapid verification at the edge control terminal and on-site execution monitoring.

10. The method for multi-machine collaborative scheduling and control of an automated warehouse based on digital twins as described in claim 9, characterized in that: The process of fusing and denoising based on execution feedback and rollback flags to form a continuously optimized scheduling loop involves the following steps: First, based on execution feedback and rollback flags, the 3D layout of the automated warehouse and multi-machine sensor data are fused and denoised to obtain a cleaned execution dataset. Then, based on the cleaned execution dataset, comparative analysis is used to assess and locate health boundary offsets, token default distribution, and time slot congestion hotspots, generating adjustment suggestions for the healthy feasible domain. Based on the adjustment recommendations, online updates are performed and the unified healthy feasible domain scheduling weights are optimized synchronously to form a continuously optimized scheduling loop.

Images